diff mbox

ARM: ux500: avoid warning in ux500_read_asicid

Message ID 2475820.dXNmmAuTaP@wuerfel (mailing list archive)
State New, archived
Headers show

Commit Message

Arnd Bergmann May 31, 2013, 8:47 p.m. UTC
phys_addr_t may be 64 bit, which causes this harmless warning
in ux500_read_asicid:

arch/arm/mach-ux500/id.c: In function 'ux500_read_asicid':
arch/arm/include/asm/io.h:159:19: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
 #define IOMEM(x) ((void __force __iomem *)(x))
                   ^
arch/arm/mach-ux500/id.c:40:9: note: in expansion of macro 'readl'
  return readl(IOMEM(UX500_VIRT_ROM + (addr & 0xfff)));

We can solve this in a nicer way by making UX500_VIRT_ROM have a proper
type to start with and calculating the address using pointer arithmetic.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Linus Walleij <linus.walleij@linaro.org>

Comments

Linus Walleij June 4, 2013, 9:23 a.m. UTC | #1
On Fri, May 31, 2013 at 10:47 PM, Arnd Bergmann <arnd@arndb.de> wrote:

> phys_addr_t may be 64 bit, which causes this harmless warning
> in ux500_read_asicid:
>
> arch/arm/mach-ux500/id.c: In function 'ux500_read_asicid':
> arch/arm/include/asm/io.h:159:19: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
>  #define IOMEM(x) ((void __force __iomem *)(x))
>                    ^
> arch/arm/mach-ux500/id.c:40:9: note: in expansion of macro 'readl'
>   return readl(IOMEM(UX500_VIRT_ROM + (addr & 0xfff)));
>
> We can solve this in a nicer way by making UX500_VIRT_ROM have a proper
> type to start with and calculating the address using pointer arithmetic.
>
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> Cc: Linus Walleij <linus.walleij@linaro.org>

Patch applied to ux500-core for v3.11, as it doesn't seem to
be a dangerous regression.

You'll get this back by pull request soon-ish.

Yours,
Linus Walleij
Arnd Bergmann June 4, 2013, 9:49 a.m. UTC | #2
On Tuesday 04 June 2013 11:23:48 Linus Walleij wrote:
> On Fri, May 31, 2013 at 10:47 PM, Arnd Bergmann <arnd@arndb.de> wrote:
> 
> > phys_addr_t may be 64 bit, which causes this harmless warning
> > in ux500_read_asicid:
> >
> 
> Patch applied to ux500-core for v3.11, as it doesn't seem to
> be a dangerous regression.
> 
> You'll get this back by pull request soon-ish.

Ok, thanks!

	Arnd
diff mbox

Patch

diff --git a/arch/arm/mach-ux500/db8500-regs.h b/arch/arm/mach-ux500/db8500-regs.h
index b2d7a0b..bb4cc41 100644
--- a/arch/arm/mach-ux500/db8500-regs.h
+++ b/arch/arm/mach-ux500/db8500-regs.h
@@ -184,7 +184,7 @@ 
 #define U8500_IO_VIRTUAL	0xf0000000
 #define U8500_IO_PHYSICAL	0xa0000000
 /* This is where we map in the ROM to check ASIC IDs */
-#define UX500_VIRT_ROM		0xf0000000
+#define UX500_VIRT_ROM		IOMEM(0xf0000000)
 
 /* This macro is used in assembly, so no cast */
 #define IO_ADDRESS(x)           \
diff --git a/arch/arm/mach-ux500/id.c b/arch/arm/mach-ux500/id.c
index 0d33d1a..392f2fd 100644
--- a/arch/arm/mach-ux500/id.c
+++ b/arch/arm/mach-ux500/id.c
@@ -21,11 +21,11 @@ 
 
 struct dbx500_asic_id dbx500_id;
 
-static unsigned int ux500_read_asicid(phys_addr_t addr)
+static unsigned int __init ux500_read_asicid(phys_addr_t addr)
 {
 	phys_addr_t base = addr & ~0xfff;
 	struct map_desc desc = {
-		.virtual	= UX500_VIRT_ROM,
+		.virtual	= (unsigned long)UX500_VIRT_ROM,
 		.pfn		= __phys_to_pfn(base),
 		.length		= SZ_16K,
 		.type		= MT_DEVICE,
@@ -37,7 +37,7 @@  static unsigned int ux500_read_asicid(phys_addr_t addr)
 	local_flush_tlb_all();
 	flush_cache_all();
 
-	return readl(IOMEM(UX500_VIRT_ROM + (addr & 0xfff)));
+	return readl(UX500_VIRT_ROM + (addr & 0xfff));
 }
 
 static void ux500_print_soc_info(unsigned int asicid)